Models & routing

Fusion

Reads the request: a classifier (TypeSafe Jev) labels its task type and complexity, and that label picks the model and reasoning effort. Naming this id sends a truncated copy of your prompt to the classifier.

Call itExamples read your key from MINIROUTER_KEY.

Fusion request
curl https://api.minirouter.sh/v1/chat/completions \
  -H "Authorization: Bearer $MINIROUTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "minirouter/fusion",
  "messages": [
    {
      "role": "user",
      "content": "Explain an API gateway in one sentence."
    }
  ],
  "max_tokens": 256
}'

How it picksAllowed on your key, within your cost tier, healthy, and able to take the request.

A classifier labels the request. The label picks the list, best first.

LabelModels
Code · Simpledeepseek/deepseek-v4.1-flash-1017 max, anthropic/claude-haiku-5.5 medium, minimax/minimax-m3.1-flash
Code · Everydayopenai/gpt-6.1-terra high, anthropic/claude-sonnet-5.5 high, meta/muse-ember max
Code · Complexanthropic/claude-sonnet-5.5 xhigh, openai/gpt-6-astra medium, meta/muse-wonder-1.0 max
Code · Expertanthropic/claude-fable-5.2 high, openai/gpt-6-astra medium, meta/muse-ember max
Reasoning · Simpledeepseek/deepseek-v4.1-flash-1017 max, anthropic/claude-haiku-5.5 medium, minimax/minimax-m3.1-flash
Reasoning · Everydayopenai/gpt-6.1-terra high, anthropic/claude-sonnet-5.5 high, meta/muse-ember max
Reasoning · Complexgoogle/gemini-4-argon high, openai/gpt-6.1-sol high, anthropic/claude-sonnet-5.5 xhigh
Reasoning · Expertanthropic/claude-fable-5.2 high, openai/gpt-6-astra medium, google/gemini-4-argon high
Writing · Simpledeepseek/deepseek-v4.1-flash-1017 max, anthropic/claude-haiku-5.5 medium, minimax/minimax-m3.1-flash
Writing · Everydayopenai/gpt-6.1-sol medium, anthropic/claude-sonnet-5.5 high, meta/muse-ember max
Writing · Complexopenai/gpt-6.1-sol medium, anthropic/claude-sonnet-5.5 high, meta/muse-ember max
Writing · Expertopenai/gpt-6.1-sol medium, anthropic/claude-fable-5.1 medium, meta/muse-ember max
Conversation · Simplemeta/muse-ember xhigh, deepseek/deepseek-v4.1-flash-1017 none, openai/gpt-6.1-terra none
Conversation · Everydayopenai/gpt-6.1-sol medium, meta/muse-ember max, anthropic/claude-fable-5.1 medium
Conversation · Complexopenai/gpt-6.1-sol medium, anthropic/claude-fable-5.1 medium, meta/muse-ember max
Conversation · Expertopenai/gpt-6.1-sol medium, anthropic/claude-fable-5.1 medium, meta/muse-ember max
No labelopenai/gpt-6.1-sol medium, meta/muse-ember max, anthropic/claude-fable-5.1 mediumUsed when the classifier cannot answer.

Live availability: Auto models.

Pricing

  • PriceNamed in x-minirouter-model. You pay for the model that answered.Provider cost plus 10%.
  • BoundThe dearest model in the list, at this router's rate.At most $11.00 in and $55.00 out per million tokens.

Notes

Full guide: Fusion.